测试视频预测报错IndexError：数组索引需为整数/布尔类型求助

问题场景 #

在为测试视频生成模型预测时触发以下错误：

IndexError: arrays used as indices must be of integer (or boolean) type

实现逻辑：

• 创建两个空列表存储预测标签与真实标签
• 遍历测试集中的每个视频，提取帧存入temp文件夹（每次迭代清空文件夹）
• 读取temp文件夹中的帧，用预训练模型提取特征，预测标签后取众数作为视频标签存入列表，同时存入真实标签

完整代码：

# creating two lists to store predicted and actual tags
predict = []
actual = []

# for loop to extract frames from each test video
count = 0
for i in tqdm(range(test_videos.shape[0])):
    videoFile = test_videos[i]
    cap = cv2.VideoCapture(test_videos[i])   # capturing the video from the given path
    frameRate = cap.get(cv2.CAP_PROP_FPS) #frame rate
    x=1
    # removing all other files from the temp folder
    files = glob('E:/FCI Bio/GP/Dataset/Paper/3/3.1/OHP/Labeled_Dataset/videos/dataset/test/temp/*')
    for f in files:
        os.remove(f)
    while(cap.isOpened()):
        frameId = cap.get(1) #current frame number
        ret, frame = cap.read()
        if (ret != True):
            break
        if (frameId % math.floor(frameRate) == 0):
            # storing the frames of this particular video in temp folder
            filename ='E:/FCI Bio/GP/Dataset/Paper/3/3.1/OHP/Labeled_Dataset/videos/dataset/test/temp/' + os.path.basename(videoFile).split('_')[0] + "_frame%d.jpg" % count;count+=1
            cv2.imwrite(filename, frame)
    cap.release()
    
    # reading all the frames from temp folder
    images = glob("E:/FCI Bio/GP/Dataset/Paper/3/3.1/OHP/Labeled_Dataset/videos/dataset/test/temp/*.jpg")
    
    prediction_images = []
    for i in range(len(images)):
        img = image.load_img(images[i], target_size=(224,224,3))
        img = image.img_to_array(img)
        img = img/255
        prediction_images.append(img)
        
    # converting all the frames for a test video into numpy array
    prediction_images = np.array(prediction_images)
    # extracting features using pre-trained model
    prediction_images = base_model.predict(prediction_images)
    # converting features in one dimensional array
    prediction_images = prediction_images.reshape(prediction_images.shape[0], 7*7*512)
    # predicting tags for each array
    prediction = model.predict(prediction_images)
    # appending the mode of predictions in predict list to assign the 
    tag to the video
    predict.append(y.columns.values[s.mode(prediction)[0][0]])
    # appending the actual tag of the video
    actual.append(os.path.basename(videoFile).split('_')[0])

错误栈：

IndexError                                Traceback (most recent call last)
~\AppData\Local\Temp\ipykernel_4364\3325358529.py in <module>
     44     prediction = model.predict(prediction_images)
     45     # appending the mode of predictions in predict list to assign the tag to the video
---> 46     predict.append(y.columns.values[s.mode(prediction)[0][0]])
     47     # appending the actual tag of the video
     48     actual.append(os.path.basename(videoFile).split('_')[0])

IndexError: arrays used as indices must be of integer (or boolean) type

错误原因 #

错误出在第46行predict.append(y.columns.values[s.mode(prediction)[0][0]])，核心问题：

1. model.predict(prediction_images)返回的是浮点数类型的概率分布数组（每个元素对应标签的预测概率），不是整数标签索引
2. 直接对概率数组取众数无意义：浮点数概率几乎不会完全重复，且众数结果仍是浮点数，无法作为索引去访问y.columns.values（索引必须是整数或布尔类型）

修复方案 #

先将每帧的概率数组转换为整数标签索引，再对这些索引取众数，最后用索引获取标签名称：

修改后的代码片段 #

替换原代码中从prediction = model.predict(prediction_images)到predict.append(...)的部分：

# predicting tags for each array (返回概率分布)
prediction = model.predict(prediction_images)
# 将每帧的概率转换为整数标签索引（取概率最大的标签）
frame_labels = np.argmax(prediction, axis=1)
# 对所有帧的标签索引取众数
video_label_idx = s.mode(frame_labels)[0][0]
# 根据索引获取对应的标签名称并加入列表
predict.append(y.columns.values[video_label_idx])
# appending the actual tag of the video
actual.append(os.path.basename(videoFile).split('_')[0])

关键说明 #

• np.argmax(prediction, axis=1)：对每个样本的概率数组，返回概率最大的标签的索引（整数类型），解决了索引类型不匹配的问题
• 对整数类型的标签索引取众数，才是真正意义上的“多数帧预测的标签”，符合视频标签的逻辑

内容的提问来源于stack exchange，提问作者Mustafa Barkat